I am currently working on a project in Python that involves using OpenCV for preprocessing user-taken images before passing them to an OCR (Optical Character Recognition) system. These images may have varying quality issues, such as poor lighting and other similar problems.
As a beginner in image processing, I would like to implement an adaptive preprocessing approach, where the preprocessing techniques employed would dynamically adjust based on the specific characteristics of each image. For example, if an image exhibits low brightness or high levels of noise, I would like to apply specific adjustments or filters accordingly.
I am wondering if there are any established techniques or best practices in Python and OpenCV that could help me achieve this adaptive preprocessing functionality. Are there any libraries or functions specifically designed for handling such scenarios? Additionally, any insights or suggestions on the overall approach would be greatly appreciated, as well as academic articles or books suggestions.
For now I am using:
-Adaptive Gaussian Thresholding
-Median Blur
-Noise removal
Probably I am gonna add something to sharpen the image.
It seems that it is not improving the quality of the data taken by the OCR.
Thank you in advance for any guidance you can provide!
Top comments (0)